As AI speeds up legal work, clients press law firms for discounts

AI efficiency creates a pricing dispute

Large US law firms are eager to demonstrate their use of AI, according to The New York Times DealBook report published on September 26. After OpenAI introduced a version of its latest model for lawyers, Sullivan & Cromwell unveiled Agreement Analyzer for reviewing deals. Cooley launched Go Public to support the S-1 drafting process before an IPO, while Ropes & Gray is developing a tool that it says can produce a detailed diligence report in hours rather than weeks.

That speed sits uneasily with hourly billing. Firms argue that AI does not always save as much time as clients assume, at least not yet. They say the technology is changing the mix of work instead: lawyers spend less time on routine tasks and more on substantive analysis.

Clients seek discounts of 20% to 30%

Jennifer Leonard, founder of Creative Lawyers, said some clients use AI agents to identify invoice entries where technology could have saved hours, then strike those items or demand a discount. Kyle Poe, vice president of legal innovation at Legora, told DealBook that in-house legal teams often ask firms to reduce the entire bill by 20% to 30%.

AI-native rivals are adding pressure. An affiliated law firm of financial-services-focused Norm AI uses upfront pricing. Norm AI raised $260 million in July at a $1.2 billion valuation from Khosla Ventures, Blackstone and Bain Capital, and now counts those investors among its clients.

The billable hour is unlikely to disappear

Of the 55 large law firms surveyed by Citi last year, 35% expected to change their billable-hour models because of AI within a year, while 65% expected changes by 2035. There is little immediate financial pressure: Wells Fargo said average revenue at large firms rose more than 12% in the first half of the year, while standard billing rates increased by double digits.

Hourly fees still distribute risk in complex disputes and transactions whose workload cannot be predicted. A hybrid model is therefore more likely: repeatable, AI-heavy work can move to fixed or project pricing, while strategy, negotiation and advocacy remain hourly. Firms are preparing for that shift, but many are still reluctant to be the first major player to lead it.